SPOILERS BELOW!!!

library(ggplot2)
library(readr)
library(dplyr)
library(gridExtra)
library(GGally)
library(gganimate)
part1 = read_csv("https://raw.githubusercontent.com/hyperc54/data-puzzles-assets/master/visualisation/chal1/chal_visu_1_q1_data.csv")

The given data are 2-dimensional points with labels x and y. There appeared to be a lot of overplotting near (0, 0), so I zoomed in there, uncovering the word “leafy.” This was the first answer.

part1 %>% ggplot(aes(x, y)) + geom_point(alpha = 0.1) + theme_bw() + ggtitle("Visualization With Transparency")

part1 %>% filter(0 < x, x < 1500, -300 < y, y < 0) %>% ggplot(aes(x, y)) + geom_point() + theme_bw() + ggtitle("Visualization of Central Points")

part2 = read_csv("https://raw.githubusercontent.com/hyperc54/data-puzzles-assets/master/visualisation/chal1/chal_visu_1_q2_data.csv")

The second data set has seven features, labeled 0 through 6. I ran pairwise correlations and plotted the resulting 2d plots, discovering a prancing unicorn. “unicorn” was the second answer.

ggpairs(part2)

part3 = read_csv("https://raw.githubusercontent.com/hyperc54/data-puzzles-assets/master/visualisation/chal1/chal_visu_1_q3_data.csv")

The third data set contains numbers labeled x and y as well as a third “mystery feature.” I visualized the x and y coordinates spatially as in part 1 and colored the points according to their “mystery value” but this did not reveal anything obvious. I thought about combining two or more coordinates and tried to look for pairwise correlations in the combined values. I thought I saw some things resembling letters in the plot of mystery_feature + y vs mystery_feature.

part3 %>% ggplot(aes(x = x, y = y, color = mystery_feature)) + geom_point()

part3 = part3 %>% mutate(x_plus = x + mystery_feature, x_minus = x - mystery_feature, y_plus = y + mystery_feature, y_minus = y - mystery_feature)
ggpairs(part3)

part3 %>% ggplot(aes(mystery_feature, y + mystery_feature)) + geom_point()

Then I thought about treating the mystery_feature value as a frame and viewing the x-y coordinates as an animation. Doing this created a stream of what looked like letters; slowing it down revealed the message containing the final message: “THE ANSWER IS CRUSH”.

anim = part3 %>% ggplot(aes(x, y)) + geom_point() + transition_time(mystery_feature)
animate(anim, end_pause = 25)

Using the three words together maps you to the final answer! (I won’t spoil this one.)